100,
re: QCOM: cdma2000/3GSM odds and ends
<< What do you mean by end to end and not just the radio interface. >>
In the case of GSM I mean the whole technology platform consisting of the GSM core network, the Base Station Subsystem, the Mobile Station (Mobile Equipment/SIM, system services, and all underlying protocols.
Terminology has changed a bit in 3G so in UMTS we have the evolved core network upgradable to an all IP network, and the UTRAN on the infrastructure side and the User Equipment (Mobile Station/UIM), and of course system services and the underlying protocols for all.
Here's a classic concise overview of the end to end GSM platform:
ccnga.uwaterloo.ca
Here is a reasonably concise overview of the end to end 3GSM (UMTS) platform:
unet.univie.ac.at
Grahame Lynch had this to say on this topic when commenting on our friend George Gilder's WSJ comments when AWS chose the GSM migration path:
Gilder forgets that there is a big difference between an air interface and a standard. CDMA is a superior air interface to TDMA, but cdmaOne is not a superior standard to GSM. This is why the TDMA fraternity is frantically attempting to converge its back-end systems with GSM, which has the lead when it comes to roaming tables and databases, messaging protocols, billing procedures and subscriber identity modules. These standards enable GSM operators to earn lots of margin-heavy extra cash.
GSM was not designed as a this and a that (an ANSI-41 core that evolved from analog days into a network with an IS-154/IS-136 TDMA or a IS-95 CDMA air interface), it was designed as an end to end digital technology whole, and the same applies to 3GSM.
When Ericsson and Nokia came to separate similar strategic decisions in early 1997 and decided to ally with Asia by approaching DoCoMo, the compromise made by DoCoMo was that their version of broadband CDMA would combine with the GSM core (evolved to accommodate it).
<< Since we both agree that wcdma will make up at least 75% of the market do you not feel Q can dominate the wcdma ASIC market in addition to dominating cdma2000 therefore becoming the wireless gorilla. >>
I believe that it will be very difficult for QCT to dominate the WCDMA silicon/processor market. It is a highly competitive market and they are going up against some major players, many of them who started development earlier, and many of whom are well along with requisite multi-mode chip development.
I think I'll wait to see how this develops. It's a little early to tell and I haven't fully adjusted to the fact that SpinCo is NoSpinCo yet. My intention was to look at this a little closer after SpinCo spun, but SpinCo didn't spin.
It would be a royalty play anyway, I would think, since WCDMA is based on a committe-based architecture, unless some proprietary (open) QCT thingie becomes the differentiator.
<< Sometimes, in my mind, I compare Qualcomm to Apple, Jacobs to Jobs. I would agree with this if you convince me that cdma will not be the dominate wireless technology at some future date. >>
I'm not even going to attempt to do that, other than to say that comittee-based architecture will dominate wireless tomorrow as it does today, that I don't think that Qualcomm's proprietary open architecture will exceed 20% of the market at such time as "CDMA" becomes the dominant radio interface in wireless (2008 or so - and yes, last year I said 2007), and Qualcomm's technology just didn't get picked to be the dominant technology by the ."committee" or (so far)the technology adopters that foot the bill.
<< Who pray tell are these companies you state will dominate wireless and why. >>
I addressed that in a post to Mike this morning, but in addition to Qualcomm we have Ericsson, Sony Ericsson, Nokia, Motorola, Samsung, Texas Instruments, Intel, Infineon, CSCO, SUN etc. as major players. I don't suspect that any single one will dominate wireless but several will dominate their respective segments.
I'll address a bit of the why (relative to one dominant company) in my upcoming Nokia Project Hunt Report. I haven't done enough DD on the others to really comment intelligently.
<< what software companies dominate wireless now or in the future. Will not Brew put Q in this category? Will Java win out? >>
One I'm interested in is Openwave. Another I already hold is SEBL. BREW is interesting and I think to the degree that Java will run under it will be successful. I think that Java (Java Card and J2ME) will be very big and dominate development environments).
<< What is GSM phase 2?? >>
In order to launch GSM in our lifetime the GSM MoU split GSM development into 2 phases.
GSM Phase one launched in 1992.
GSM Phase 2 launched in early 1995 and featured enhanced data services and an enhanced vocoder that doubled GSM voice capacity.
About the same time GSM Phase 2+ development began, starting with SIN Toolkit, including HSCSD, and now finalizing with GPRS. Soon it will be on to 3G, which uses GPRS as a bearer service.
<< The majority of handsets sold through at least 2006 will be GSM phase 2 or GSM phase 2+ (GSM/GPRS). ... Why do you say 2006. Docomo has already started service >>
Mass deployment of WCDMA won't begin outside of Japan till 2004 or so. At this time the majority of handsets being sold will be GSM and probably 80% of those will be GSM Phase 2+ (GPRS) by that time. It will take some time for GSM users to migrate to WCDMA (not much different than the AMPS to CDMA migration or the AMPS to TDMA migration).
<< China is building a 15 million subscriber system in less then 1 yr.
This is what I'm not understanding. >>
China is building out a $2 Billion + 15 million subscriber line second generation cdmaOne network (upgradable to 1xRTT) and (will be) trialing 1xRTT. 1xRTT was not fully commercialized at the time they commenced buildout (still isn't) and even now so far as I know there are no dual-mode GSM/1xRTT handsets.
The vast majority of infra spending last year and continuing into this year was on GSM/GPRS networks or GPRS upgrades to existing GSM networks. Somewhere I have numbers for last year. Haven't pinned em down for H1 this year. If I had to guess (using Yankee and Nokia forecasts) this year the market is about $100 Billion and cdma will do maybe 20 Billion with GSM/GPRS doing probably $55 Million or so.
<< So if wcdma infra is beginning to ramp now why will it not begin to ramp in mass in 4th Qtr 02 (one yr from sample shipments)and why will it take 6 yrs to build out. >>
Limited Cap Ex, and lack of experience and business justifications with packet data and for wireless data ... not to mention compelling content and devices.
The fact that GPRS is a year late didn't help things. GSM carriers will spend the next year (and perhaps beyond) primarily focused on stabilizing and improving the packet data aspects of their newly launched GPRS networks.
Worse yet for us Qualcommers, right now the GSM carriers that are licensed for 3G are basically saying GPRS will be adequate to meet customer needs for the foreseeable future.
Bottom line is that wireless data is a technology push and there is no compelling consumer (market) demand.
<< Does not the fact that Q is now taking a more active/positive role in the 3gpp wcdma specification writings move up the time tables and give Q an edge in dominating wcdma asic sales?? >>
I don't really see it as moving up timetables appreciably. If Qualcomm intends to dominate wcdma ASICS they have to take an active role. If you watch standards evolve you can kind of tell that when a company submits a flurry of contributions (clarifications, corrections, recommendations, etc.) for the first time, it is a signal that their developer's adrenalin is flowing and they are well along in development. Its a positive sign to be sure.
<< this very disappointing time of Qualcomm >>
It is really not that disappointing a time for Qualcomm per se, although wireless sure has sucked wind for awhile, and so have techs. There is lots of good in store for Qualcomm and its investors. Heck we got a nice little + 7.18 % bounce today, but let's think longer term. There is the better part of a year for wireless to work its way out of the hole as we really start the transition to data, although when market conditions improve so will wireless and Qualcomm should be out in front.
As always JMHO or the facts as I interpret them, with an opinion or 2 tossed in for good measure.
- Eric - |
